Background: Within the context of students' self-regulated learning, the interplay between learners' individual characteristics and the context of testing have been emphasized for assessing learning outcomes. Aims: The present study examined metacognitive monitoring and control processes in elementary schoolchildren's test taking behaviour and explored the impacts of these metacognitive skills for the accuracy and the quantity of test performance. Sample and methods: A total of 133 participants from third and fifth grade did a cloze test about a previously learned science topic, gave confidence judgments for every answer, and were then allowed to cross-out answers if they wished. Two different mock scoring schemes for test performance were compared with a control group. Results: Results revealed well-developed monitoring skills indicating that by the age of 9 children can reliably distinguish between correct and incorrect answers. As for control skills, 11- and 12-year-olds proved to be better able to improve their test performance by selectively withdrawing answers that would have been incorrect than the 9- to 10-year-olds. Conclusions: The study offers evidence for the impact of metacognitive processes in students' learning outcomes and documents strategic behaviour during test taking, as well as developmental progression in the involved skills.